EDP Sciences
Free Access
Issue
A&A
Volume 555, July 2013
Article Number A34
Number of page(s) 14
Section Planets and planetary systems
DOI https://doi.org/10.1051/0004-6361/201220414
Published online 25 June 2013

© ESO, 2013

1. Introduction

Comets are small bodies formed at the birth of the solar system 4.6 billion years ago. Since they did not evolve much, they are potential witnesses of the physical and chemical processes at play at the beginning of our solar system (Ehrenfreund & Charnley 2000). Their status of “fossils” gives them a unique role to understand the origins of the solar system, not only from the physical and dynamical points of view but also from the chemical one (thanks to the knowledge of the compounds of the nucleus). When the comet gets closer to the Sun, the ices of the nucleus sublimate to form the coma (the comet atmosphere) where oxygen atoms are detected. Oxygen is an important element in the chemistry of the solar system given its abundance and its presence in many molecules including H2O, which constitutes 80% of the cometary ices. Oxygen atoms are produced by the photodissociation of molecules coming from the sublimation of the cometary ices. Chemical reactions Eqs. (1) to (5) involve possible parent molecules (Bhardwaj & Raghuram 2012; Festou & Feldman 1981):

thumbnail Fig. 1

Energy-level diagram for [OI] lines.

Open with DEXTER

Oxygen atoms have been detected in comets through the three forbidden lines observed in emission at 5577.339 Å (the green line) and at 6300.304 Å and 6363.776 Å (the red lines) (Swings 1962). These lines come from the deexcitation of the upper state (2p4)1S0 to the (2p4)1D2 state for the 5577.339 Å line and from the (2p4)1D2 state to the (2p4)3P1,2 states for the doublet red lines (see Fig. 1). The lifetime of the oxygen in the 1D state is ~110 s which is much longer than the lifetime of ~1 s for the O(1S). The measurement of the green line is more difficult owing to its fainter intensity and the many C2 lines located around it. One of the first theoretical studies was carried out by Festou & Feldman (1981). They reviewed the production rate of O(1S) and O(1D) obtained in the laboratory from H2O, CO, and CO2 photodissociations and measured the corresponding 1S/1D ratio. Recently, Bhardwaj & Raghuram (2012) made a new model for oxygen atom emissions and calculated the production rate of O(1S) and O(1D) for chemical reactions (1) to (5). The estimated 1S/1D ratios are significantly different from Festou & Feldman (1981) and are given in Table 1.

Table 1

Emission rates and 1S/1D ratios relevant to the quiet Sun obtained by Bhardwaj & Raghuram (2012) for photodissociation reactions at 1 AU.

Up to now, little systematic researchwork has been done to study these lines at various heliocentric distances because the detection of the forbidden lines requires both high spectral and high spatial resolutions. Morrison et al. (1997) observed oxygen in comet C/1996 B2 (Hyakutake) with the 1-m Ritter Observatory telescope, Cochran & Cochran (2001), and Cochran (2008) analyzed [OI] lines in the spectra of eight comets observed at the McDonald Observatory (see Bhardwaj & Raghuram 2012 for a complete review of these measurements).The present paper reports the results obtained for a homogeneous set of high-quality spectra of 12 comets of various origins observed since 2003 with the UVES spectrograph mounted on the 8-m Kueyen telescope of the ESO VLT.

First, we measured the intensities of the three forbidden oxygen lines. Two ratios were evaluated: the ratio between the two red lines, I6300/I6364, and the ratio of the green line to the sum of the red lines, I5577/(I6300 + I6364) which we denote as G/R hereafter. The purpose of the latter is to determine the main parent molecule of the oxygen atoms by comparing our results with the Bhardwaj & Raghuram (2012) effective excitation rates. We also followed how the G/R ratio depends on the heliocentric distance by analyzing spectra of comets C/2001 Q4 (NEAT) and C/2009 P1 (Garradd) at small and large distances from the Sun.

Finally, we measured the full width at half maximum (FWHM) of the three lines. There is a longstanding debate on the FWHM measurement because the first observations made by Cochran (2008) contradict the theory. The intrinsic width of the green line is wider than the red ones, while the theory predicts the opposite (Festou 1981).

In this paper, we present the analysis of the three forbidden oxygen lines in various comets at different heliocentric distances. The observation of all the comets in our sample are explained in Sect. 2. In Sects. 3 and 4, details about data reduction and data analysis are described. Results of the two ratios and the FWHM of the [OI] lines are given in Sect. 5 followed by discussion in Sect. 6.

2. Observations

Our analysis is based on 12 comets listed in Table 2, considering the two components of 73P/Schwassmann-Wachmann 3 (B and C) as two comets. This sample is characterized by a wide diversity. These comets have various dynamical origins (external, new, Jupiter family, Halley type) and are observed at different heliocentric distances (from 0.68 AU to 3.73 AU). Some are split comets like 73P or did a close approach to Earth. Others have been observed at large heliocentric distances such as C/2001 Q4 and C/2009 P1. The observing material is made of a selection of 48 high signal-to-noise spectra obtained during the last ten years (Manfroid et al. 2009; Jehin et al. 2009, and references therein) with the cross-dispersed echelle spectrograph UVES (Dekker et al. 2000). A dichroic filter splits the light beam into two arms (a red one from 3000 to 5000 Å and a blue one from 4200 to 11 000 Å). In most observations, a slit of 0.45′′ × 11′′ was used. With such a slit, the resolving power is R = λλ = 110   000 in the red arm where the three [OI] lines are observed. All spectra were recorded with nearly the same instrumental setting. Such high resolution is needed to isolate the [OI] emission lines from the telluric lines and other cometary lines (C2, NH2) (see Fig. 2). The slit was always centered on the comet nucleus. For a comet located at 1 AU from the Earth, the spatial area covered by the slit is approximatively 320 km by 8000 km. The size of the slit for each individual spectrum is given in Table 3.

Table 2

Orbital characteristics of the 12 comets and their classification.

thumbnail Fig. 2

The 6300.304 Å, 6363.776 Å and 5577.339 Å oxygen lines (telluric (⊕) and cometary) in a 3600 s spectrum of comet 8P/Tuttle obtained in March 2008. Thanks to the high resolution of the UVES spectrograph, the cometary and telluric lines are well separated when the Doppler shift is larger than 15 km s-1. These spectra are representative of the quality of the sample.

Open with DEXTER

3. Data reduction

The 2D spectra were reduced with the UVES pipeline reduction program. For data obtained until 2008, the UVES pipeline (version 2.8.0) was used within the ESO-MIDAS environment to extract, for each order separately, the 2D spectra, bias-subtracted, flat-fielded, and wavelength-calibrated. ESO-MIDAS is a special software containing packages to reduce ESO data1. For a given setting, the orders were then merged using a weighting scheme to correct for the blaze function, computed from high S/N master flat-fields reduced in exactly the same way. This procedure leads to a good order merging. The post-2008 observations were reduced with the gasgano CPL2 interface (version 2.4.3) of the refurbished UVES pipeline that directly provides accurately merged 2D spectra. The 1D spectra were extracted by averaging the 2D ones, with cosmic rejection. Standard stars were similarly reduced and combined to derive instrumental response functions used to correct the 1D spectra. The wavelength calibration was computed using thorium argon (Th-Ar) spectra. The calibration revealed small position shifts in the lines because the Th-Ar spectra were usually taken in the day and not right after the comet observations. This shift was removed using nine cometary lines of NH2 in the vicinity of the red [OI] lines for which laboratory wavelengths are well known.

Before doing any measurement, the telluric absorption lines were removed and the solar continuum contribution subtracted using the BASS3 and Kurucz4 spectra. These two spectra are solar spectra, with atmospheric absorption lines for the BASS spectrum and without for the Kurucz one (2005). Around the red doublet line at 6300 Å, there are indeed telluric absorption lines mostly due to O2 molecules in the Earth’s atmosphere, which could lead to an underestimate of the forbidden oxygen line intensity (see Fig. 3).

The comet solar continuum has three contributions. First and foremost is the one from the reflection of sunlight on the dust particles in the coma. Another one appears when the observations are made in or close to the twilight. A third contribution is, in a few cases, the background radiation by the Moon. To remove these contaminations, we Doppler-shifted the BASS spectrum to the proper values and scaled the intensities until the solar features are completely removed. This provides the final spectrum, such as the one presented in Fig. 4 for comet 8P/Tuttle, free of any solar and telluric lines.

4. Data analysis

After the reduction and the correction of the data, we measured both the intensity and the FWHM of the three forbidden oxygen lines by fitting a Gaussian profile using the IRAF5 software. The observed width is the convolution of the instrumental profile with the natural width: (7)where the instrumental width corresponds to the width of the Th-Ar lines.

These measurements can only be performed with the best accuracy when the cometary oxygen lines are separated from the telluric [OI] lines well, i.e, when the geocentric velocitiy of the comet exceeds 15 km s-1. Using the deblend function of the splot package in IRAF, we could measure the [OI] lines for Doppler shifts as low as 7 km s-1.

5. Results

We measured the two ratios and the FWHM of the three [OI] lines. To reach such goals, a relative calibration using the well known UVES instrumental response curve provided by ESO was enough. Nevertheless, we investigate how to produce absolute flux-calibrated spectra in order to derive production rates of the oxygen parent species in a future work. Therefore, in this current study, the production rates of H2O to assess the activity of the comet have been taken from the literature.

5.1. Intensity ratios

The two measured intensity ratios, I6300/I6364 and G/R = I5577/(I6300 + I6364), are displayed for each spectrum in Figs. 5 and 6 respectively. When the collisional quenching is neglected, the intensity of a line is given by (Festou & Feldman 1981) (8)where τp is the photodissociative lifetime of the parent, α the yield of photo-dissociation, β corresponds to the branching ratio for the transition, and N is the column density of the parent. Table 4 lists the line intensities I (in ADU) and the ratios for all the spectra. Table 5 shows the standard deviation (σ) of the ratio for all the spectra of a given comet. Since usually most of the spectra of the same comet are obtained during a short interval of time at similar heliocentric and geocentric distances, as well as the same exposure time, the standard deviation of these measurements provides a good estimate of the error. Indeed, this error includes the photon noise and errors coming from the telluric absorption and solar continuum corrections.

The red doublet ratio is remarkably similar for all comets. The average value over the whole sample of comets is 3.11   ±   0.10. The error corresponds to the standard deviation (σ). The ratio agrees with the value of 3.09   ±   0.12 obtained by Cochran (2008) from a sample of eight comets but with better accuracy. Since both red lines are transitions from the same level to the ground state, τp, α, and N are equal in Eq. (8), and the intensity ratio I6300/I6364 is equivalent to the branching ratio β6300/β6364. Our results are indeed in very good agreement with the theoretical value of the branching ratio of 3.096 computed for the quantum mechanics (Galavis et al. 1997). Storey & Zeippen (2000) have published a new theoretical value of 2.997 taking relativistic effects into account, which should be an improvement over previous determinations. Higher value derived from cometary spectra could point to a systematic error due to a small blend of the 6300.304 Å [OI] line with an unidentified cometary feature. To check that these systematics are present, we first took the average red line ratio of the three comets at large heliocentric distances (C/2001 Q4, C/2003 K4, and C/2009 P1) since most of the fluorescence lines and then their contamination disappear far from the Sun. We obtained an average ratio of 3.13 ± 0.07. Second, we computed the same ratio, but this time for the terrestrial nightglow, and found a ratio equal to 3.20   ±    0.05 obtained again at the greatest cometary distances. These measurements agree better with the Galavis et al. (1997) value, but the error are too large to discard Storey & Zeippen (2000) so we cannot exclude a systematic error.

The G/R ratio has an average value of 0.11   ±   0.07 for the complete sample (see Fig. 6).This result agrees with the value of 0.09   ±   0.01 found by Cochran (2008). The dispersion is higher because it includes comets at large heliocentric distances that have different G/R ratios. If we only take comets at <2.5 AU into account, the average and dispersion are equal to 0.09 ± 0.02. This leads to the conclusion that H2O is the main parent molecule producing oxygen atoms according to Bhardwaj & Raghuram (2012) values (Table 1).

Table 3

Individual spectra.

To see whether the quenching could play a role in our results, we computed the G/R ratio with respect to the water production rates (Fig. 7). The water production rates of each comet has been compiled from the literature (cf. Table 8). Since those comets have been observed under different conditions (of production rates and heliocentric distances), the lack of trend in this analysis suggests that quenching is negligible and does not significantly affect our main conclusions. However, we have started to work on a model based on the Bhardwaj & Raghuram (2012) paper and on Monte Carlo simulations to estimate the quenching. A careful computation of the quenching is not an easy task, and we will give our new results, including a detailed discussion of the quenching, in a forthcoming paper.

5.2. Line widths

Table 6 lists the intrinsic line velocity widths (FWHM(v) (km s-1)). The FWHM(v) is obtained from the FWHMintrinsic(λ) given in the Eq. (7) using the relation (9)where λ[OI] corresponds to the wavelength of the considered oxygen line, and c is the speed of light (km s-1). The error on the [OI] line widths of each comet is given by the standard deviation of the N spectra (cf. Table 7). This error is about 5% for the red lines and 10% for the green line, which is fainter. The two red line widths are equal within the errors (1.61 ± 0.34 km s-1 for the 6300 Å line and 1.56 ± 0.54 km s-1 for the 6364 Å line), which is not a surprise since both lines are transitions from the O(1D) state to the ground state (Fig. 8). This result is also consistent with the values of Cochran (2008) (1.22 ± 0.36 km s-1). Our mean value of the [OI] cometary green line is wider than the red lines and equal to 2.44 ± 0.28 km s-1. This peculiarity has already been noticed by Cochran (2008), who found a mean velocity of 2.49 ± 0.36 km s-1, in good agreement with our value. Figures 911 present the width of the three [OI] cometary lines, the three [OI] telluric lines, and the width of some representative neighboring cometary lines for comparison: NH2 lines in the red region and C2 lines in the green region. The intrinsic average widths of the cometary NH2 and C2 lines are, respectively, 1.00 ± 0.20 km s-1 and 1.28 ± 0.17 km s-1 and correspond well to what is expected for the gas expansion velocity in the coma at 1 AU. The ejection velocity due to the extra energy coming from the photodissociation of the parent molecule should be represented as the excess velocity that results from the subtraction of the NH2 or C2 cometary lines to the [OI] lines. Considering this, we find an average ejection velocities equal to 0.48 ± 0.16 km s-1 for the red lines and 1.17 ± 0.29 km s-1 for the green line. This analysis and Fig. 12 show clearly that the [OI] cometary green line is wider than the red lines. This width could be explained if the excess energy for the formation of the O(1S) state is greater than for the O(1D) state. Contrary to what is observed, theoretical models using Ly-α photons as excitation source give an excess velocity of 1.6 km s-1 for the O(1S) state and a value of 1.8 km s-1 for O(1D) in the case of water photodissociation (Festou 1981).

thumbnail Fig. 3

The telluric absorption line correction in the vicinity of the red [OI] lines. The black spectrum has been corrected (thick line), the red one is uncorrected.

Open with DEXTER

thumbnail Fig. 4

Subtraction of the solar continuum. Upper spectrum (red) shows the spectrum of comet 8P/Tuttle not corrected for solar lines while in the lower one (black), solar continuum has been removed. Measurements of [OI] are made with the lower spectrum.

Open with DEXTER

thumbnail Fig. 5

The red doublet ratio for each comet. The spectra of a given comet have the same symbols. The average value of the sample (3.11 ± 0.10) is shown by the solid line. The dashed lines show the standard deviation (σ).

Open with DEXTER

thumbnail Fig. 6

Same as Fig. 5 for the G/R ratio. The mean value is 0.11 ± 0.07. Note the data points for comet Q4 (NEAT) at 3.7 AU.

Open with DEXTER

In Figs. 13 and 14, we compare the velocity of the three oxygen lines and the H2O production rate from the literature (cf. Table 8). We find that the velocity of the green oxygen line slightly increases with the water production rate. In Figs. 15 and 16, we provide similar plots for the velocities of C2 and NH2. The results agree with Tseng et al. (2007) but the error bars are large. Indeed, since the resolution of the spectrometer is at the limit for measuring the line broadening, we cannot improve their relation.

6. Discussion

6.1. The G/R ratio at large heliocentric distance

In Fig. 6, we noticed that C/2001 Q4 (NEAT) has a G/R ratio of about 0.3, so clearly much higher than other comets with values around 0.1. C/2001 Q4 was at 3.7 AU from the Sun, and this peculiarity led us to look more carefully at comets observed at large heliocentric distances (see Fig. 17 and Decock et al. 2011).

If we only consider the data taken at r < 2 AU, the G/R ratio average value is equal to 0.09 ± 0.02, which is in good agreement with the ratio obtained by Bhardwaj & Raghuram (2012) and Festou & Feldman (1981) for H2O as the parent molecule (see Table 1). The high values of G/R in Q4 (NEAT) at 3.7 AU could be explained by the increasing contribution, at large heliocentric distances, of other parent molecules producing oxygen atoms. It is well known that the sublimation of H2O significantly decreases beyond 3 AU, while the sublimation of other ices like CO and/or CO2 dominates comet activity (i.e., Crovisier & Encrenaz2000). To investigate this hypothesis, we observed comet C/2009 P1 (Garradd) at four different heliocentric distances from 3.25 AU to 2.07 AU. Figure 18 presents the G/R intensity ratios as a function of r for all spectra. A relation between the heliocentric distance and the G/R intensity ratio can be seen: the latter is getting rapidly larger when the comet gets farther from the Sun. We added in this plot the Cochran & Cochran (2001) and Cochran (2008) average values coming from their sample of eight comets. All these values are clumped around 1 AU with a ratio of ~0.1. C/2001 Q4 is one of their eight comets and its G/R ratio is 0.09   ±   0.02 at r = 0.98 AU. A few other measurements found in the literature for comets observed at heliocentric distances beyond 2 AU were also added to the graph. Furusho et al. (2006) used the Subaru telescope to analyze the forbidden oxygen lines in comet 116P/Wild 4 observed at r   =   2.4 AU. Capria et al. (2010) measured the intensity of these lines during and after the outburst of comet 17P/Holmes, which occurred on October 24, 2007 when the comet was located at about r = 2.5 AU. McKay et al. (2012b) studied the [OI] lines of the two comets C/2006 W3 Christensen and C/2007 Q3 Siding Spring, respectively at 3.13 and 2.96 AU, with the ARCES echelle spectrometer of 3.5-m telescope at Apache Point Observatory. While the quality of the data is not as good, because the lines are faint and sometimes heavily blended with the telluric line, all these measurements show a relatively high value of the G/R ratio when r > 3 AU. This confirms the hypothesis that oxygen is also coming from other molecules. Both CO and/or CO2 are obvious candidates as they produce high values of the G/R ratio (see Table 1).

Table 4

Intensities given in ADU (analog-to-digital units) and measured line ratios for the three forbidden oxygen lines in the spectra of each comet.

Table 5

Average values of the line ratios for each comet.

thumbnail Fig. 7

The G/R ratio against the water production rate.

Open with DEXTER

Using HST/STIS6 observations, Feldman et al. (2004) estimated at 4% the CO abundance relative to water in comet C/2001 Q4 located at 1 AU from the Sun. The study made by Biver et al. (2012) on comet C/2009 P1 provided an average value of 5% for the CO abundance when the comet was at 1.9 AU at pre-perihelion phase (October 2011). At large heliocentric distances, CO/H2O had to be even higher, so the CO molecule could be a candidate to explain such a high G/R ratio. However, Bhardwaj & Raghuram (2012) show that after the photodissociation of H2O, the next main source for the green line emission is CO2 with 10–40% abundance relative to water. They compared their model with the observational values obtained for C/1996 B2 (Hyakutake) by Morrison et al. (1997) and Cochran (2008) when the comet was around 1 AU. Hyakutake is also rich in CO, and its abundance was evaluated at 22% at 1 AU (Biver et al. 1999). Assuming a CO2 abundance of 1% and an O(1S) yield of 0.2%, Bhardwaj & Raghuram (2012) concludes that the production rates of the O(1S) are similar for CO and CO2 (~10–25%) with such abundances. Therefore, since C/2001Q4 and C/2009 P1 have less CO than Hyakutake, CO2 molecules could instead be the main contributor to the formation of the green line emission for these two comets observed at large heliocentric distance. The CO2 measurements are unfortunately difficult to obtain to confirm this hypothesis. If this is the case, estimated G/R ratios might be used as a proxy for the CO2 relative abundance, but such a relation should be carefully calibrated on a sample of comets with known CO2.

Using Eq. (8) and only considering the main parent species of oxygen atoms (i.e H2O, CO2, and CO), we can write the G/R ratio as (10)}where corresponds to the effective production rates of the O(1S) and O(1D) states7. However, if we consider that oxygen atoms only come from H2O and CO molecules, we found that G/R ratio does not change strongly: it varies from 0.08 to 0.1 for a CO abundance between 10% and 80%. Therefore, we assume that the [OI] atoms are only produced by the H2O and CO2 photodissociations. Equation (10) can thus be simplified as (11)Therefore, the CO2/H2O abundance can be computed by (12)The are independent of the heliocentric distance because all of them depend on the same heliocentric distance. This is also the case for the values given in Table 1. Considering the same values for and βgreen, we also computed the CO2/H2O abundance from the G/R ratios of comets C/2001 Q4 and C/2009 P1 measured by Cochran (2008) and McKay et al. (2012a), respectively. These results are shown in Fig. 19. We noticed that CO2/H2O is as high as 75% in C/2001 Q4 and that CO2 starts to contribute to the G/R ratio at 2.5 AU in the comets. This heliocentric distance limit is also shown by Ootsubo et al. (2012) from a sample of 17 comets observed with the AKARI IR space telescope.

Table 6

Measured observed FWHM (Å) and intrinsic velocity widths (km s-1) for the three forbidden oxygen lines in the spectra of each comet.

Therefore, the forbidden oxygen line measurement could become a new way to determine the CO2 abundance in comets at different heliocentric distances from ground observations, while the direct measurement of CO2 molecules is only possible from space.

6.2. The [OI] lines widths

One explanation for the large width of the [OI] green line could be that exciting photons producing oxygen atoms in the 1S state have more energy than the Lyman α photons and/or come from another parent molecule. Indeed, Fig. 7 of Bhardwaj & Raghuram (2012) shows that the wavelength region for the major production of O(1S) coming from CO2 seems to be the 955–1165 Å band. New theoretical studies could be necessary to evaluate the excess velocity for the O(1S) state coming from the photodissociations of both H2O and CO2.

In Fig. 12, we also notice that the [OI] red lines of the comet C/2001 Q4 are wider than in other comets, while the width of the green line is normal compared to other comets. This peculiarity seems real as both red lines show the same width, and they are 2-σ away from the average value of the sample (1.58 ± 0.30 km s-1). The telluric lines in C/2001 Q4 have the usual width (0.99 ± 0.10 km s-1 for the 6300 Å line, 0.87 ± 0.04 km s-1 for the 6364 Å line and 0.80 ± 0.26 km s-1 for the green line), excluding a problem with the spectra or the analysis. The width of the red lines and the green line is similar in only this case (2.57 ± 0.16 km s-1 for the 6300 Å line, 2.49 ± 0.15 km s-1 for the 6364 Å and 2.46 ± 0.13 km s-1 for the 5577 Å line) and could give us clues to the process in play. As previously discussed, at large heliocentric distances, the [OI] lines could mainly come from the CO2 molecules which mainly photo-dissociate in the 1S state, while at low heliocentric distances, they essentially come from H2O, which primarily dissociates in the 1D state (Table 1). This means that at large distances, the red lines are mainly produced through the 1S-1D channel, together with the green line, while at low distances they are mostly produced directly from the 1D state since H2O dominates. We thus expect the widths of the red and green lines to be similar at high distances because the O(1S) and O(1D) atoms are mostly produced from a same molecule, while they can differ at low distances, as observed. CO2 is the best candidate, and the larger width could be explained by the main excitation source of CO2, the 955–1165 Å band, which is more energetic than Ly-α photons, as shown in Fig. 11 of Bhardwaj & Raghuram (2012). This particularity is not seen for comet C/2009 P1 (Garradd), maybe because the spectrum is not taken at a sufficiently large heliocentric distance, and the CO2 abundance was lower than in C/2001 Q4 (NEAT). Either way, this hypothesis needs to be confirmed by observing other comets at large heliocentric distances (>3.5 AU).

Table 7

Average values of the intrinsic line width for each comet.

thumbnail Fig. 8

The 6300.304 Å FWHM(v) against the 6363.776 Å one. The different spectra for a given comet are denoted with the same symbol.

Open with DEXTER

thumbnail Fig. 9

The FWHM(v) of the 6300.304 Å [OI] cometary line (■), the [OI] telluric line (▴), and nearby NH2 cometary lines (●) (km s-1) for one spectrum per comet. The typical errors correspond to the standard deviation (σ) obtained considering all the spectra of comet 8P/Tuttle. We have also checked this error by measuring σ randomly for a few spectra of other comets. For C/2001 Q4, the NH2 cometary lines are missing because no fluorescence lines were detected at such a large heliocentric distance.

Open with DEXTER

thumbnail Fig. 10

Same as Fig. 9 for the 6363.776 Å line.

Open with DEXTER

6.3. 103P/Hartley 2

The Jupiter family comet 103P/Hartley 2 has been found to be poor in CO (Weaver et al. 2011) and rich in CO2 with an CO2 abundance of ~20% relative to the water from EPOXI measurements (A’Hearn et al. 2011) and with an abundance of ~10% from the ISO observations (Crovisier et al. 1999). Thanks to Eq. (12), we could evaluate the CO2 abundances from our G/R ratios measured at ~1 AU and compare them with the values obtained by EPOXI and ISO observations. The results are given in Table 9. The mean G/R ratio of 0.09   ±   0.01 is similar to those of other comets observed below 2.5 AU, and it gives a relative abundance of CO2 of only ~3%, while a G/R ratio of ~0.1 to 0.2 would be needed to reach a CO2 abundance as high as ~10% to 20%.

thumbnail Fig. 11

Same as Fig. 9 for the 5577.339 Å line and with C2 cometary lines.

Open with DEXTER

thumbnail Fig. 12

The 5577.339 Å line width against the average width of the red lines. The different spectra of a given comet are denoted by the same symbol. The green line is always larger than the red lines except for C/2001 Q4.

Open with DEXTER

We do not confirm a high value of CO2 abundance for 103P/Hartley 2 as claimed by McKay et al. (2013) or a large dispersion in the abundances. Their values might be higher because they used a of Bhardwaj & Raghuram (2012) equal to 2.6 × 10-8 s-1, while the value provided in Table 1 of this paper is 6.4 × 10-8 s-1.

thumbnail Fig. 13

Average of the [OI] red lines FWHM versus the water production rate.

Open with DEXTER

thumbnail Fig. 14

Same as Fig. 13 but for the 5577.339 Å line.

Open with DEXTER

The discrepancy with the EPOXI and ISO observations CO2 abundance is interesting. Below 2.5 AU, G/R values are distributed from 0.05 for comets like 9P/Tempel 1 and 8P/Tuttle and can go up to 0.12 for C/2002 T7 (LINEAR). The spread from comet to comet is higher than the errors estimated from the dispersion from the spectra of a same comet, which could be explained by an intrinsic variation in G/R from comet to comet due to the content of CO2. The G/R range, on the order of 0.1 (implying a CO2 variation of ~20%), is close to the range of CO2 abundance in comets of various origins measured below 2.5 AU (A’Hearn 2012).

Some comets, such as 9P/Tempel 1 and 8P/Tuttle, have lower values than the pure water case (giving a G/Rmin = 0.08 from Bhardwaj & Raghuram (2012)) and that the 103P/Harley 2 G/R ratio is too low compared to EPOXI CO2 abundances could indicate a problem with the models. The pure water ratio should be 0.05 or lower based on the comets with the lowest values. 9P/Tempel 1 has a CO2 abundance of ~7% (Feaga et al. (2007a), Feaga et al. (2007b). Bhardwaj & Raghuram (2012) tried values from 0% to 1% and finally chose 1%, but this value is uncertain. Without CO2, G/R. If we use 0.5% for , G/Rmin is equal to 0.04 instead of 0.08, which would be in better agreement with our lowest values.

6.4. Deep impact

The spectra of comet 9P/Tempel 1 were analyzed before and after the July 4 collision with the Deep Impact spacecraft. No difference was observed in the intensity ratios and the lines widths before and after the impact (Fig. 20). We have no spectrum during the impact. The first spectrum after the impact was taken at 2 453 559.954 UT (i.e. ~6 h after the impact). Cochran (2008) did not notice any change during the impact either.

6.5. Solar activity

Our data were obtained over the 23d solar cycle characterized by its maximum of activity in 2001 and its minimum activity in 2008. Despite the decrease in the solar activity during the observations of C/2001 V1 (NEAT) and 8P/Tuttle, no change in the intensities and widths of lines were noticed. The solar activity does not appear to have any influence on the results of our work.

thumbnail Fig. 15

Same as Fig. 13 but for the NH2 line.

Open with DEXTER

thumbnail Fig. 16

Same as Fig. 13 but for the C2 line.

Open with DEXTER

thumbnail Fig. 17

Two hours spectra of comet C/2001 Q4 obtained with the ESO VLT. These spectra are the most distant detection of [OI] lines in a comet (at 3.7 AU). The lines are well detected and well separated from the telluric line. The width of the green and the red lines are the same.

Open with DEXTER

thumbnail Fig. 18

The G/R intensity ratio as a function of the heliocentric distance (AU). The same symbol is used for the different spectra of each comet. The most distant comets have higher ratios. Other values from the literature are also plotted. The solid line represents the average value only for our sample of spectra, and the standard deviation (σ) range is denoted with horizontal dashed lines. If we only consider the data at r < 2 AU, the average value of the G/R ratio is 0.09 ± 0.02. The vertical dashed line at 3 AU represents the distance out to which the sublimation of water is strongly decreasing (Crovisier & Encrenaz 2000).

Open with DEXTER

thumbnail Fig. 19

Evolution of the CO2/H2O abundance as a function of the heliocentric distance (AU). Assuming that a higher G/R ratio comes from a more important contribution of CO2 for large heliocentric distance comets, and measuring this ratio can give us the CO2 abundance from ground observations.

Open with DEXTER

thumbnail Fig. 20

The G/R intensity ratio of comet 9P/Tempel 1 from JD 2 453 553.955 to 2 453 562.956 UT. The Deep Impact event corresponds to the vertical line (53 559 UT). The average value is shown by the solid red line and the dashed lines correspond to the standard deviation (σ). The mean value around the impact (from 2 453 553.955 UT to 2 453 562.956 UT) is 0.04   ±   0.01. No variation of the ratio was noticed after the impact.

Open with DEXTER

Table 8

Water production rates and their references.

Table 9

The CO2/H2O abundances estimated from Eq. (12) for the comet 103P/Hartley 2.

7. Conclusion

From January 2003 to September 2011, 12 comets of various origins were observed at different heliocentric distances, and 48 high-resolution spectra were obtained with the UVES spectrograph at VLT (ESO). Within this whole sample, we observed the three [OI] forbidden oxygen lines with high resolution and signal-to-noise and we measured the intensity ratios (I6300/I6364 and G/R = I5577/(I6300 + I6364)), as well as the FWHM (v) of the lines. The results can be summarized as follows.

  • 1.

    The I6300/I6364 ratio (3.11 ± 0.10) is in very good agreement with the branching ratio obtained by quantum mechanics, especially the work done by Galavis et al. (1997). However, some systematic errors on the order of 0.1 are possible so that we cannot exclude the new value of Storey & Zeippen (2000).

  • 2.

    From theoretical values given in Table 1, the G/R ratio for comets observed below 2 AU (0.09 ± 0.02) confirms H2O as the main parent molecule photodissociating to produce oxygen atoms. However, when the comet is located at greater heliocentric distances (r ≥ 2.5 AU), the ratio increases rapidly, showing that another parent molecule is contributing. We have shown that CO2 is the best candidate. Measuring the G/R ratios could then be a new way to estimate the abundances of CO2, a very difficult task from the ground. Assuming that only the photo-dissociation of H2O and CO2 produce [OI], we found a relation between CO2 abundance and the heliocentric distance of the comet. The C/2001 Q4 (NEAT) abundance of CO2 at 3.7 AU is found to be ~75%.

  • 3.

    The intrinsic green line width is wider than the red ones by about 1 km s-1. Theoretical estimations considering Ly-α as the only excitation source for the two states led to the conclusion that the excess energy for O(1D) is greater than for O(1S), which contradicts our observations. This discrepancy might be explained by the different nature of the excitation source and/or a contribution of CO2 as parent molecule to the O(1S) state. Indeed, Bhardwaj & Raghuram (2012) have shown that for the photodissociation of CO2, the main excitation source might instead be the 995–1165 Å band. To check this hypothesis quantitatively, it would be necessary to theoretically estimate the excess energy for the oxygen atoms when the wavelength band is 995–1165 Å, accounting for the photodissociation of both H2O and CO2. The widths of the three [OI] lines are similar in C/2001 Q4 at ~3.7 AU. This could agree with CO2 being the main contributor for the three [OI] lines at a large heliocentric distance. Other comets at large r have to be observed to test this hypothesis.

  • 4.

    More CO2 (and CO) abundance determinations, together with G/R oxygen ratios and line widths at different heliocentric distances, are clearly needed in order to give a general conclusion about the oxygen parent molecule.

  • 5.

    The CO2 rich comet 103P/Hartley 2 (~20%) does not present a high G/R ratio normally expected from Eq. (12). We suggest a new value of 0.5% for that could also explain the low values of G/R obtained for comets 9P/Tempel 1 and 8P/Tuttle.


5

IRAF is a tool for the reduction and the analysis of astronomical data (http://iraf.noao.edu).

6

HST/STIS is the acronym for Hubble Space Telescope/Space Telescope Imaging Spectrograph (http://www.stsci.edu/hst/stis).

7

The production rates are given in Tables 1 and 2 of Bhardwaj & Raghuram (2012) (i.e s-1, s-1, s-1, s-1, s-1 and s-1) and βgreen = 0.91 (Slanger et al. 2006).

Acknowledgments

A.D. acknowledges the support of the Belgian National Science Foundation F.R.I.A., Fonds pour la formation à la Recherche dans l’Industrie et l’Agriculture. E.J. is Research Associate FNRS, D.H. is Senior Research Associate at the FNRS and J.M. is Research Director of the FNRS. C. Arpigny is acknowledged for the helpful discussions and constructive comments.

References

All Tables

Table 1

Emission rates and 1S/1D ratios relevant to the quiet Sun obtained by Bhardwaj & Raghuram (2012) for photodissociation reactions at 1 AU.

Table 2

Orbital characteristics of the 12 comets and their classification.

Table 3

Individual spectra.

Table 4

Intensities given in ADU (analog-to-digital units) and measured line ratios for the three forbidden oxygen lines in the spectra of each comet.

Table 5

Average values of the line ratios for each comet.

Table 6

Measured observed FWHM (Å) and intrinsic velocity widths (km s-1) for the three forbidden oxygen lines in the spectra of each comet.

Table 7

Average values of the intrinsic line width for each comet.

Table 8

Water production rates and their references.

Table 9

The CO2/H2O abundances estimated from Eq. (12) for the comet 103P/Hartley 2.

All Figures

thumbnail Fig. 1

Energy-level diagram for [OI] lines.

Open with DEXTER
In the text
thumbnail Fig. 2

The 6300.304 Å, 6363.776 Å and 5577.339 Å oxygen lines (telluric (⊕) and cometary) in a 3600 s spectrum of comet 8P/Tuttle obtained in March 2008. Thanks to the high resolution of the UVES spectrograph, the cometary and telluric lines are well separated when the Doppler shift is larger than 15 km s-1. These spectra are representative of the quality of the sample.

Open with DEXTER
In the text
thumbnail Fig. 3

The telluric absorption line correction in the vicinity of the red [OI] lines. The black spectrum has been corrected (thick line), the red one is uncorrected.

Open with DEXTER
In the text
thumbnail Fig. 4

Subtraction of the solar continuum. Upper spectrum (red) shows the spectrum of comet 8P/Tuttle not corrected for solar lines while in the lower one (black), solar continuum has been removed. Measurements of [OI] are made with the lower spectrum.

Open with DEXTER
In the text
thumbnail Fig. 5

The red doublet ratio for each comet. The spectra of a given comet have the same symbols. The average value of the sample (3.11 ± 0.10) is shown by the solid line. The dashed lines show the standard deviation (σ).

Open with DEXTER
In the text
thumbnail Fig. 6

Same as Fig. 5 for the G/R ratio. The mean value is 0.11 ± 0.07. Note the data points for comet Q4 (NEAT) at 3.7 AU.

Open with DEXTER
In the text
thumbnail Fig. 7

The G/R ratio against the water production rate.

Open with DEXTER
In the text
thumbnail Fig. 8

The 6300.304 Å FWHM(v) against the 6363.776 Å one. The different spectra for a given comet are denoted with the same symbol.

Open with DEXTER
In the text
thumbnail Fig. 9

The FWHM(v) of the 6300.304 Å [OI] cometary line (■), the [OI] telluric line (▴), and nearby NH2 cometary lines (●) (km s-1) for one spectrum per comet. The typical errors correspond to the standard deviation (σ) obtained considering all the spectra of comet 8P/Tuttle. We have also checked this error by measuring σ randomly for a few spectra of other comets. For C/2001 Q4, the NH2 cometary lines are missing because no fluorescence lines were detected at such a large heliocentric distance.

Open with DEXTER
In the text
thumbnail Fig. 10

Same as Fig. 9 for the 6363.776 Å line.

Open with DEXTER
In the text
thumbnail Fig. 11

Same as Fig. 9 for the 5577.339 Å line and with C2 cometary lines.

Open with DEXTER
In the text
thumbnail Fig. 12

The 5577.339 Å line width against the average width of the red lines. The different spectra of a given comet are denoted by the same symbol. The green line is always larger than the red lines except for C/2001 Q4.

Open with DEXTER
In the text
thumbnail Fig. 13

Average of the [OI] red lines FWHM versus the water production rate.

Open with DEXTER
In the text
thumbnail Fig. 14

Same as Fig. 13 but for the 5577.339 Å line.

Open with DEXTER
In the text
thumbnail Fig. 15

Same as Fig. 13 but for the NH2 line.

Open with DEXTER
In the text
thumbnail Fig. 16

Same as Fig. 13 but for the C2 line.

Open with DEXTER
In the text
thumbnail Fig. 17

Two hours spectra of comet C/2001 Q4 obtained with the ESO VLT. These spectra are the most distant detection of [OI] lines in a comet (at 3.7 AU). The lines are well detected and well separated from the telluric line. The width of the green and the red lines are the same.

Open with DEXTER
In the text
thumbnail Fig. 18

The G/R intensity ratio as a function of the heliocentric distance (AU). The same symbol is used for the different spectra of each comet. The most distant comets have higher ratios. Other values from the literature are also plotted. The solid line represents the average value only for our sample of spectra, and the standard deviation (σ) range is denoted with horizontal dashed lines. If we only consider the data at r < 2 AU, the average value of the G/R ratio is 0.09 ± 0.02. The vertical dashed line at 3 AU represents the distance out to which the sublimation of water is strongly decreasing (Crovisier & Encrenaz 2000).

Open with DEXTER
In the text
thumbnail Fig. 19

Evolution of the CO2/H2O abundance as a function of the heliocentric distance (AU). Assuming that a higher G/R ratio comes from a more important contribution of CO2 for large heliocentric distance comets, and measuring this ratio can give us the CO2 abundance from ground observations.

Open with DEXTER
In the text
thumbnail Fig. 20

The G/R intensity ratio of comet 9P/Tempel 1 from JD 2 453 553.955 to 2 453 562.956 UT. The Deep Impact event corresponds to the vertical line (53 559 UT). The average value is shown by the solid red line and the dashed lines correspond to the standard deviation (σ). The mean value around the impact (from 2 453 553.955 UT to 2 453 562.956 UT) is 0.04   ±   0.01. No variation of the ratio was noticed after the impact.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.